HNHacker News
TopNewBestAskShowJobs

spencerchubb

1,275 karma · joined December 3, 2021

https://twitter.com/spencer_chubb
submissionscomments
spencerchubb··on Ask HN: Kodak, VW, Intel – Why big companies fail?
companies undergo natural selection much like organisms do. the difference is companies aren't made of cells, they're made of people. it's a higher level of abstraction.

in the case of kodak, their cameras were simply worse than competitor's cameras. either kodak must die, or it must kill parts of its current product and make a better product. to give a biological analogy, it would have to cut off a limb and regrow the limb to stay viable. it's very complicated and risky to do that kind of surgery

spencerchubb··on Show HN: I made crowdwave – imagine Twitter/Reddit but every post is a voicemail
your word choice sounds quite negative: illiterate, disengaged, minimum effort

why is it a bad thing to consume media that best aligns with how humans perceive information?

spencerchubb··on Show HN: I made crowdwave – imagine Twitter/Reddit but every post is a voicemail
and now my Twitter feed is dominated by video

maybe it was a matter of cost. over time the cost of storage has gone down, and people are more engaged by richer forms of media. video > photo > text

spencerchubb··on Llama 3.1 Omni Model
Sounds like it's trained on LJ Speech dataset, which is one of the best datasets and very commonly used
spencerchubb··on A bottle of water per email: the hidden environmental costs of using AI chatbots
I argue it's impossible to measure the total environmental impact of anything, and it's a waste of time to even try.

Let's break this down. There is water used at inference time and training time. The humans working on the project consume water. The building in which they did the project uses water. The whole supply chain for the computers? Good luck measuring the water usage in there

This may seem pedantic, but I promise there is a point to this.

Measuring environmental impact is like trying to understand a neural network.

If you want to discourage water usage, the only way is to tax the marginal water used at any step in the supply chain. You don't need to know the total water usage for the last step of the supply chain, in this case, an LLM

spencerchubb··on Knocking out one key gene leads to autistic traits
Would you agree that a cure should be available? Obviously everyone can choose for themselves if they want it cured.
spencerchubb··on Knocking out one key gene leads to autistic traits
No. Searching for an autism cure does not bring stigma to autistic people, the same way searching for a cancer cure does not bring stigma to cancer patients.

Searching for a cure is just trying to improve people's quality of life. If a cure were hypothetically found, nobody would be forced to get it. If someone likes having autism, more power to them!

spencerchubb··on Knocking out one key gene leads to autistic traits
Who said anything about eugenics or killing autistic people? The comment chain is discussing hypothetical cures for autism
spencerchubb··on How to succeed in MrBeast production (Leaked PDF)
I read the entire document and I don't understand where you saw bad culture or micromanaging.

Some people may not like the fact that they pull all nighters, but that's a matter of opinion. Clearly some people do like the terms of employment, otherwise they wouldn't work there.

spencerchubb··on Twelve sentenced for violent home invasion robberies to steal cryptocurrency
Multisig wallet where you and your spouse both have to sign. In case one dies, the multisig can expire after a certain amount of time and let you access with just one key

Of course this assumes you have a spouse

spencerchubb··on How America's universities became debt factories
Sure there may competition among suppliers. When you artificially inflate demand, the price will still go up. One does not negate the other
spencerchubb··on How America's universities became debt factories
Unless I missed it in the article, I don't think the author mentioned one of the biggest factors. The big lenders in America are effectively extensions of the US government, and the government guarantees over 90% of loans.

That means there is no risk (for the lender) to give out student loans. The taxpayers take on the risk

spencerchubb··on Show HN: Meet.hn – Meet the Hacker News community in your city
There are multiple cities in the U.S. with the name of my city, so unfortunately this doesn't work for me.
spencerchubb··on Breaking Down OnlyFans' Economics
Just don't tell the consumer that it's Gen AI
spencerchubb··on Kolmogorov-Arnold networks may make neural networks more understandable
We know the process to train a model, but when a model makes a prediction we don't know exactly "how" it predicts the way it does.

We can use the economy as an analogy. No single person really understands the whole supply chain. But we know that each person in the supply chain is trying to maximize their own profit, and that ultimately delivers goods and services to a consumer.

spencerchubb··on Why Copilot Is Making Programmers Worse at Programming
People have said that types make people worse at programming too lol
spencerchubb··on Why Copilot Is Making Programmers Worse at Programming
Solutions from any source are often wrong. Stack Overflow, intellisense, human peers. The question you have to ask is: does it make you more productive even though there are mistakes?
spencerchubb··on Google Illuminate: Books and papers turned into audio
You left this out

"The transformer processes the entire sequence all at once by using something called self attention"

spencerchubb··on Google Illuminate: Books and papers turned into audio
How do you define great?
spencerchubb··on Building the same app using various web frameworks
End-to-end correctness does not matter in a comparison. If every app implements the same functionality, then that is a valid comparison. Even if that functionality is incorrect.
spencerchubb··on Building the same app using various web frameworks
You're focusing on the wrong aspect of the post. This post is comparing web frameworks, not showing how to build a correct csv handling app.
spencerchubb··on Why Not Comments
My team comments way too much. I constantly see stuff like this

def fetch_data(comment_id):

  Args:  
    - comment_id: The id of the comment to fetch.

  Returns: The comment data
  
# Fetch comment data

data = fetchCommentData()

spencerchubb··on ESPN AI recap of Alex Morgan’s final professional match fails to mention her
isn't the point of seo to get in front of consumers?
spencerchubb··on Jd – JSON Diff and Patch
usr stands for user system resources
spencerchubb··on Manual 'Till It Hurts
It goes to show how horrible javascript is that it has the concept of a "build step".
spencerchubb··on Manual 'Till It Hurts
you shouldn't have to copy paste html. any backend language you use probably has templates
spencerchubb··on Minifying HTML for GPT-4o: Remove all the HTML tags
it is theoretically possible, but the results and bandwidth would be worse. sending an image that large would take a lot longer than sending text
spencerchubb··on Llms.txt
I didn't say that humans are second-class to LLMs. Nor does the proposal suggest that. It's an additional mode in addition to the webpage that humans use
spencerchubb··on Reflection 70B, the top open-source model
I wonder how good it is with multi-turn conversations
spencerchubb··on Llms.txt
It's recognizing that the needs of a human are different from the needs of an LLM.
← PreviousPage 4 of 16Next →